conftrace_

safety-utility trade-off

5 papers

Explore in graph

Co-occurring keywords · papers mentioning both

large language model (2) role-playing dialogue (1) adaptive dynamic multi-preference (1) jailbreak prevention (1) inference-time search (1) jailbreak vulnerability (1) attention head analysis (1) indirect prompt injection (1) dialogue agent (1) activation steering (1)

Papers

ARGUS: Defending Against Multimodal Indirect Prompt Injection via Steering Instruction-Following Behavior CVPR 2026 EcoAlign: An Economically Rational Framework for Efficient LVLM Alignment CVPR 2026 Retrieval-Augmented Defense: Adaptive and Controllable Jailbreak Prevention for Large Language Models ACL 2026 Safety-Utility Conflicts Are Not Global: Surgical Alignment via Head-Level Diagnosis ACL 2026 The Rise of Darkness: Safety-Utility Trade-Offs in Role-Playing Dialogue Agents ACL 2025